AAV9-DOK7 Gene Therapy Reduces Disease Severity in Smn^2B/− SMA Model Mice

Abstract

Spinal Muscular Atrophy (SMA) is an autosomal recessive neuromuscular disease caused by deletions or mutations in the survival motor neuron (SMN1) gene. An important hallmark of disease progression is the pathology of neuromuscular junctions (NMJs). Affected NMJs in the SMA context exhibit delayed maturation, impaired synaptic transmission, and loss of contact between motor neurons and skeletal muscle. Protection and maintenance of NMJs remains a focal point of therapeutic strategies to treat SMA, and the recent implication of the NMJ-organizer Agrin in SMA pathology suggests additional NMJ organizing molecules may contribute. DOK7 is an NMJ organizer that functions downstream of Agrin. The potential of DOK7 as a putative therapeutic target was demonstrated by adeno-associated virus (AAV)-mediated gene therapy delivery of DOK7 in Amyotrophic Lateral Sclerosis (ALS) and Emery Dreyefuss Muscular Dystrophy (EDMD). To assess the potential of DOK7 as a disease modifier of SMA, we administered AAV-DOK7 to an intermediate mouse model of SMA. AAV9-DOK7 treatment conferred improvements in NMJ architecture and reduced muscle fiber atrophy. Additionally, these improvements resulted in a subtle reduction in phenotypic severity, evidenced by improved grip strength and an extension in survival. These findings reveal DOK7 is a novel modifier of SMA.

Introduction

Spinal Muscular Atrophy (SMA) is a devastating neurodegenerative disease and a leading genetic cause of infant mortality with a worldwide incidence of 1 in 10,000 live births¹. Without early medical intervention, most individuals with severe SMA die by two years of age and experience progressive atrophy of skeletal muscle and degeneration of alpha motor neurons². SMA arises from insufficient levels of Survival Motor Neuron (SMN) protein, an essential and ubiquitously expressed housekeeping protein that functions in a variety of RNA metabolism pathways ^3,4. Increasing SMN has proven to be the most effective therapeutic strategy and the recent approval of two breakthrough SMN-inducing modalities, Spinraza and Zolgensma, has transformed the clinical landscape of this disease^5,6. However, the specific molecular mechanisms of degeneration in SMA are still being elucidated in order to identify additional therapeutic targets for future therapies to address the breadth of the disease, non-responsive patients, and to expand the therapeutic window. These “SMN-independent” modalities have aimed to be either neuroprotective or muscle activating^7,8. In SMA, impairments in neuromuscular junctions (NMJs) include delayed maturation, reduced quantal release, decreased endplate area, and denervation^9,10. Additionally, these defects can precede other motor neuron pathologies such as axonal degeneration and motor neuron death¹¹. Thus, therapeutic correction of NMJ pathology could prevent motor neuron degeneration and preserve aspects of muscle function. This warrants further characterization of the mechanisms underlying NMJ degeneration.

Several lines of evidence suggest NMJ defects in SMA are influenced by dysfunction in the Agrin/MuSK signaling pathway, which is essential for the development, maturation, and maintenance of NMJs¹². Agrin, a motor neuron-derived proteoglycan that reinforces acetylcholine receptor (AChR) clustering at the endplate of developing NMJs, is mis-spliced in motor neurons of SMA mice¹³. Agrin transcripts in SMA mice exhibit the aberrant exclusion of exons 32–33, or “Z-exons”, which are essential for the AChR clustering property of Agrin¹⁴. Overexpression of Z⁺ Agrin increases NMJ innervation and size, reduces muscle fiber atrophy, and extends survival in SMN∆7 SMA mice¹⁵. Phenotypic improvements in SMNΔ7 mice were also observed following administration of the Agrin biological NT-1654, further confirming the importance of Agrin to the pathogenic mechanism of SMA ¹⁶.

Modification by Z⁺ Agrin implicates a novel set of putative molecular targets for SMA, as Agrin’s post-synaptic receptor complex, Lrp4 and MuSK, signal through multiple downstream factors^17–19. One such factor, downstream of tyrosine kinases 7 (DOK7), has recently gained attention as a protective modifier of several neuromuscular diseases^20,21. DOK7 is a noncatalytic scaffold protein that is essential for NMJ development, functioning in the activation of MuSK and also in signal transduction following stimulation of MuSK by Agrin²². The importance of DOK7 is emphasized by the observation that mice lacking DOK7 fail to develop NMJs and that mutations in DOK7 are associated with Limb Girdle Myasthenia^23–26. Conversely, overexpression of DOK7 enlarges NMJs and provides a significant therapeutic benefit in Emery Dreyefuss Muscular Dystrophy (EDMD) and Amyotrophic Lateral Sclerosis (ALS)^20,21,27. Administration of AAV9-DOK7 in the hSOD1-G93A mouse model of ALS increased endplate area and improved innervation of affected NMJs, suggesting that DOK7 elicits retrograde signals that influence both pre- and post-synaptic properties of NMJs²¹. Taken together, these DOK7 activities suggest a potential modifying role in the SMA context through enlargement of NMJs. To test this hypothesis, we employed AAV9-mediated gene delivery of DOK7 in the Smn^2B/− intermediate mouse model of SMA.

Results

We and others have previously shown that, compared to severe models of SMA, the slightly milder Smn^2B/− or pharmacologically-induced intermediate models are better-suited for the characterization of SMN-independent modalities^28–36. Here, we utilize the Smn^2B/− mouse model in the congenic C57BL/6J background^37,38. Although these mice are ambulatory, they possess important hallmarks of SMA disease pathology including NMJ defects, skeletal muscle atrophy, loss of motor neuron soma, and a predictably early mortality. These characteristics make the Smn^2B/− mouse a useful experimental system to assess the protective effects of DOK7 gene therapy. We constructed a self-complementary AAV9 (scAAV9) vector expressing DOK7 from the constitutive chicken beta-actin promoter to express DOK7 in Smn^2B/− mice. scAAV9-DOK7 was administered intravenously on postnatal day 1 (P1) at a dose of 1×10¹¹ vector genomes per mouse, a dose we have previously used in a variety of SMA contexts ^31,39,40. DOK7 protein levels were increased in hindlimb skeletal muscle following treatment, confirming that administration of the scAAV9-DOK7 vector induced robust upregulation of DOK7 in the target tissue (Fig 1a). Conceptually, since we were proposing that DOK7 functioned independently of Smn, it was also important to demonstrate that scAAV9-DOK7 treatment did not significantly increase SMN expression, which was confirmed by western blot (Fig 1b).

Figure 1: scAAV9-DOK7 improves phenotype of Smn^2B/− mice.

(A) Western blot of DOK7 in hindlimb skeletal muscle. (B) Western blot of SMN in hindlimb skeletal muscle. (C) Kaplan-Meier curve of Smn^2B/− mice treated with scAAV9-DOK7. Difference between scAAV9-DOK7-treated and untreated Smn^2B/− mice was calculated using log-rank Mantel-Cox test (D) Weight gain of Smn^2B/− mice treated with scAAV9-DOK7. (E) Average grip strength on P17 and P21. Groups were compared using Student’s t-test. (*P<0.05, ***P<0.001)

Smn^2B/− untreated mice lived an average of 21.2 ± 2 days, and we found that scAAV9-DOK7 prevented early deaths and induced a subtle, but statistically significant, extension in mean survival to 22.2 ± 2 days (Fig. 1C). These results are consistent with the recent report that repletion of Z⁺ Agrin in motor neurons of severe SMNΔ7 mice also extends survival^15,16. scAAV9-DOK7-treated mice did not exhibit a significant increase in weight gain compared to Smn^2B/− untreated mice (Fig. 1d), consistent with other SMN-independent strategies that extended survival, but did not impact total body weight ^31,40. To determine if treatment resulted in a functional improvement in motor activity, we employed a four-limb grip strength assay. scAAV9-DOK7 significantly improved grip strength at mid-stage disease progression compared to untreated SMA littermates and improvement was more pronounced at disease end stage (Fig. 1e). Taken together, these results demonstrate that modulation of Agrin/MuSK signaling through scAAV9-DOK7 treatment promotes modest gross phenotypic improvements, suggesting that pathogenic signaling events at affected NMJs are partially corrected.

SMA patients and mouse models exhibit denervation and reduction in motor endplate area of NMJs in a subset of skeletal muscles, resulting in a loss of connectivity in neuromuscular synaptic transmission that significantly contributes to disease pathology ^9,10,41. Because overexpression of DOK7 has been shown to increase endplate area through enhanced recruitment of AChR clustering, we assessed whether scAAV9-DOK7 treatment can suppress reduction of endplate area in SMA. In the Smn^2B/− mouse model, the NMJs at the transverse abdominis (TVA) muscle are vulnerable to degeneration¹¹. We assessed the impact of scAAV9-DOK7 on NMJs of the TVA muscle by performing immunohistochemical staining on P17, mid-symptomatic disease stage. scAAV9-DOK7 treatment prevented the reduction in endplate area compared to untreated Smn^2B/− mice, and also significantly increased endplate area compared to unaffected Smn^2B/+ controls (Fig. 2a,b). To assess whether this activity confers presynaptic improvements in NMJs, we quantified the percentage of fully, partially, and denervated endplates. scAAV9-DOK7 treatment did not increase the average percentage of innervated endplates compared to untreated Smn^2B/− mice (Fig. 2c,d). These results suggest that DOK7 improves pre- and post-synaptic aspects of degenerating NMJs in motor neuron diseases^20,21.

Figure 2: scAAV9-DOK7 improves endplate area of NMJs in Smn^2B/− mice.

Transverse abdominus muscles were dissected from mice harvested on P17. (A) Representative images of neuromuscular junction endplates labelled with α-bungarotoxin (red). Images were taken with 40X objective lens (scale bars = 20 μm). (B) Scatter plot representing quantification of motor endplate area. Bars represent mean ± SEM. (C) Representative images of neuromuscular junctions stained with neurofilament (green) and synaptophysin (green) to label the presynaptic terminals and α-bungarotoxin (red) to label endplates. (D) Bar graph representing average percentage of fully innervated, partially innervated, and denervated motor endplates. Bars represent mean ± SEM. Statistical analysis was performed comparing degree of fully-innervated endplates. Experimental groups were compared using one-way ANOVA with Tukey’s multiple comparisons test. (Smn^2B/+ n=4; Smn^2B/− n=7; Smn^2B/− + scAAV9-DOK7 n=7) (***P < 0.001, ^nsP > 0.05)

Due to the observed improvements in NMJ pathology, we hypothesized that DOK7 may reduce motor neuron degeneration. Although motor neuron cell death and NMJ pathology are believed to be regulated by separate processes, delivery of muscle-enhancing factors has previously been shown to increase the number motor neuron soma in the ventral horn of SMA mice^42–44. Towards this end, we examined the L4-L6 spinal cord at P21, a late symptomatic time point, using immunohistochemistry (Fig. 3a). As expected, untreated Smn^2B/− mice exhibited a significant loss of motor neuron soma compared to unaffected Smn^2B/+ mice, while the average count of L4-L6 soma in scAAV9-DOK7-treated mice was between the unaffected and SMA counts, however, neither comparison was statistically significant (Figure 3b). scAAV9-DOK7 did not increase motor neuron soma size, as there was no increase in soma area or perimeter compared to Smn^2B/− mice (Fig. 3c). These results suggest scAAV9-DOK7 does not improve central defects.

Figure 3: scAAV9-DOK7 does not improve central defects in Smn^2B/− mice.

(A) Motor neuron soma analyses were performed at P21 (n=3 mice). Representative images of immunofluorescent labelling of ChAT-positive (red) motor neuron soma and NeuroTrace Nissl (green) in the L4-L6 spinal cord (Scale bars = 50 μm). (B) Scatter plot representing number of ChAT-positive motor neuron soma in the L4-L6 spinal cord. (C) Scatter plot representing cross sectional area of ChAT-positive motor neuron soma in the L4-L6 spinal cord. Bars represent mean ± SEM. Experimental groups were compared using one-way ANOVA with Tukey’s multiple comparisons test. (**P < 0.01, ***P < 0.001, ^nsP > 0.05)

The loss of contact between motor neurons and skeletal muscle contributes to the atrophy of muscle fibers in SMA. Improvements in NMJ pathology following scAAV9-DOK7 gene therapy suggest that treatment also reduces the impact of skeletal muscle atrophy. To test this hypothesis, we focused our analysis on the tibialis anterior (TA) and soleus (SO) muscles, which consist predominantly of fast twitch and slow twitch fibers, respectively. Cross sectional analysis of both fibers reveals significant reduction in muscle fiber area in the untreated SMA cohort compared to the unaffected littermates. scAAV9-DOK7 significantly improved muscle fiber area of TA and SO muscles compared to the untreated Smn^2B/− mice (Fig. 4a,b). These results, together with the functional improvement in motor activity, support a muscle protective role of DOK7 in SMA.

Figure 4: scAAV9-DOK7 reduces muscle fiber pathology in Smn^2B/− mice.

Muscle analysis was performed at P17 (n=4 mice). (A) Representative images of cross sections of tibialis anterior and soleus muscles dissected on P17. Sections were stained with laminin (green). Images were taken with 40× objective lens (Scale bars = 50 μm). (B) Scatter plot representing cross sectional area of TA and SO muscle fibers. Experimental groups were compared using one-way ANOVA with Tukey’s multiple comparisons test. (***P < 0.001)

Discussion

Recent advances in SMA therapy have brought promise to the SMA community, as two highly-efficacious therapies have been approved: Spinraza, an SMN2-splice switching antisense oligonucleotide; and Zolgensma, an scAAV9-mediated gene therapy of the SMN coding sequence, which has applied for FDA approval following success in clinical trials⁶. With these breakthrough therapies, early therapeutic intervention can lessen the impact of disease severity for patients of the entire clinical spectrum of SMA. As the long-term requirements for an optimal therapeutic regimen remain to be fully determined, additional consideration must be given to SMA patients that are initiated on therapy post-symptomatically^45–47. Investigation of the molecular mechanisms underlying SMA has revealed the high degree of complexity associated with degeneration, and that an increasing number of putative therapeutic targets modify disease severity⁴⁸. This warrants further identification and characterization of novel, disease-modifying therapeutic targets.

Following recent reports that modulation of Agrin-MuSK signaling reduces severity in SMA mice, our findings suggest that the MuSK-activating protein DOK7 is a novel protective modifier of SMA. Treatment with scAAV9-DOK7 confers subtle benefits on the phenotype of intermediate Smn^2B/− SMA mice, evidenced by modest extension in survival, increased weight gain, and improved grip strength. We identified several histological improvements that likely explain the improvements in overall phenotype. scAAV9-DOK7 treatment in Smn^2B/− mice dramatically increases NMJ endplate area of NMJs of compared to both untreated Smn^2B/− mice and unaffected controls, consistent with the known function of DOK7 in AChR clustering.

There are distinctions to be made between our results and those following repletion of Z⁺ Agrin in SMA mice. Although both factors act in the same signaling pathway, overexpression of Z⁺ Agrin did not increase endplate area above that of wild type controls, unlike scAAV9-DOK7¹⁵. This can be attributed to differing roles of DOK7 and Agrin, in which DOK7 functions both to activate MuSK prior to Agrin signaling and also to propagate the MuSK signal following Agrin signaling. Thus, our results suggest that targeting of muscle-derived factors involved in NMJ organization, such as DOK7, might be better-suited to attain enlargement of motor endplates than the motor neuron-derived Agrin.

Our results contribute to the growing body of evidence that DOK7 is a potential therapeutic target that can be utilized to treat a panel of neuromuscular diseases. In the case of SMA, DOK7 could be targeted in combination with highly efficacious, SMN-inducing therapies to enhance the integrity of affected NMJs. Towards this aim, characterization of the electrophysiological impact of DOK7 upregulation in SMA is necessary to determine if the observed histological improvements correlate with functional improvements in synaptic transmission. Finally, DOK7-mediated therapy represents a putative strategy for prolonging the therapeutic window SMA treatment and it should be assessed whether DOK7-mediated therapy enhances the functionality of NMJs upon post-symptomatic administration.

Materials and Methods

Animal procedures and delivery of therapeutics

All animal experiments and conditions were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Missouri (protocol #9091).

Smn^2B/− mice of the C57/BL6 background were a gift from Dr. Rashmi Kothary at the University of Ottawa, Canada. Animals of both sexes were used in all experiments. Treatment groups were assigned at random. Genotyping was performed by PCR assay following tail biopsy on P0. 1×10¹¹ vector genomes of scAAV9-DOK7 were administered systemically via intravenous (IV) injection using the superficial vein on postnatal day 1. Grip-strength assays were conducted by placing all limbs on grid and giving mice two trials, one for training and the second for data collection (BioSeb Model BP32025, Vitrolles). Mice were fed low-fat stock diets (Harlan Teklad 8640).

Production and purification of scAAV9-DOK7 vector

scAAV9-DOK7 was constructed via triple transfection of Rep2Cap9 (a gift from James Wilson, University of Pennsylvania, Philadelphia, PA), pHelper, and scAAV-CBA-DOK7 in HEK 293T (ATCC® CRL-3216™) cells using 25-kDa polyethyleneimine. Virus purification and quantification were performed following previously used protocols^49,50.

Histological analysis and quantification of skeletal muscle fiber area

Four animals from both control cohorts and the scAAV9-DOK7-treated cohort were randomly selected and harvested at P17. Tibialis anterior (TA) and soleus (SO) muscles were dissected and flash frozen in O.C.T. media (Tissue-Tek) using liquid nitrogen-cooled isopentane. Frozen muscle was cryosectioned at 18 μm per section. Muscle fibers were labeled using anti-laminin primary antibody (1:300; catalog L9393, Sigma) and Alexa Flour 647-conjugated Donkey anti-Rabbit (1:200; Catalog AP187SA6, EMD Millipore) secondary antibody. Samples were mounted using Citiflour AFI mounting media (Ted Pella, Inc.). Imaging was performed using a laser scanning confocal microscope (20X objective; Leica TCS SP8, Leica Microsystems, Inc.). Samples were blinded and approximately 200 muscle fibers were analyzed using ImageJ software (NIH). Individual muscle fibers were pooled across mice per treatment condition. Statistical analysis was performed using a one-way ANOVA with Tukey’s multiple comparisons test⁵¹.

Immunohistochemistry of motor neuron soma

Three animals from each treatment group were randomly selected at P21, euthanized, and perfused in 4% Formaldehyde in PBS. The L4-L6 region was dissected, cross sectioned, and stained with ChAT (1:100; catalog AB144P; MilliporeSigma) and NeuroTrace Nissl (Thermo Fisher). Representative images were taken using a laser scanning confocal microscope (20X objective; Leica TCS SP8, Leica Microsystems, Inc.). Motor neuron counts and soma area were quantified using ImageJ software (NIH)⁵¹.

Immunohistochemistry of Neuromuscular Junctions

Four animals from Smn^2B/+ and seven from untreated and scAAV9-DOK7 treated Smn^2B/− cohort were randomly selected and harvested at P17. Transverse abdominus muscles were dissected and fixed in 4% Formaldehyde in PBS. Muscles were stained using anti-neurofilament-heavy (1:4000; catalog CPCA-NF-L, EnCore) and anti-synaptophysin (1:200; catalog YE269, Life Technologies™) primary antibodies, and Alexa Flour 488-conjugated Donkey anti-Chicken (1:400; EMD Millipore) and Donkey anti-Rabbit (1:200; EMD Millipore) secondary antibodies. Acetylcholine receptors were labeled with Alexa Flour 594-conjugated α-bungarotoxin (Life Technologies™). Muscles were whole mounted using Citiflour AF1 mounting media (Ted Pella, Inc.). NMJ analysis was performed on 4 randomly selected fields of view per mouse (20X objective; Leica DM5500 B, Leica Microsystems Inc.). Endplate area was measured using ImageJ software (NIH). Representative images were taken using a laser scanning confocal microscope (20X objective; Leica TCS SP8, Leica Microsystems, Inc.)⁵¹.

Immunoblot

Hindlimb skeletal muscles were harvested at P17. Tissues were lysed using JLB buffer [50 mM Tris pH=8.0, 150 mM NaCl, 10% Glycerol, 20 mM NaH₂PO₄, 50 mM NaF, 2mM EDTA, 5 mM NaVO₃] supplied with cOmplete™, Mini Protease Inhibitor Cocktail (Roche). Following SDS-PAGE, protein was transferred to Immobilon®-P transfer membrane (MilliporeSigma) and probed with the following primary antibodies: anti-SMN (1:10,000; 610647, BD Transduction Laboratories™), anti-Calnexin (1:2000; catalog C4731, Sigma), anti-DOK7 (A-7) (1:1000; catalog sc-390856, Santa Cruz Biotechnology). Horseradish peroxidase-conjugated secondary antibodies were used (Jackson ImmunoResearch Laboratories). Immunoblots were visualized using (BioSpectrum® 815 Imaging System, UVP, LLC)⁵¹.